Pure scale data recorded during this year’s summer world championships is unprecedented. FIFA, the tournament organizer, will track approximately 150 million data points per match. Inside the ball, IMU (inertial measurement unit) monitoring sensors will record 500 movements per second to track the ball’s movement.
If this sounds excessive, Patrick Lucey can go further. “The problem with football is that there are more permutations (in the game) than there are atoms in the universe,” he says.
Lucey is the Chief Scientist at Stats Perform, a data and artificial intelligence company whose work underpins almost the entire global soccer ecosystem. Their statistics are used in every aspect of the current game. It enables player tracking and multi-million player transfer fees, helps the coaching staff choose tactics and lineups, and develops procedures for corner kicks and free kicks. Players apply it to negotiate contracts, broadcasters apply it for entertainment.
Artificial intelligence is now making it possible to collect data from matches around the world like never before, and team staff are pushing the boundaries to process this data at unprecedented speeds. During the World Cup, humans and artificial intelligence will manipulate and analyze information to find novel solutions.
Teams participating in this year’s Cup will also have access to a customized AI agent Powered by Lenovo. This is FIFA’s attempt to level the playing field. Whether this is enough or not is another matter.
“The data is detailed, multi-agent and adversarial. What we do in sports is most similar to autonomous vehicles – we look at trajectories,” says Lucey. “If you think about one team, there are 10 factorial permutations in terms of player order. If you include the opposition, it just explodes.”
Even smaller countries have found novel ways to apply technology. Curaçao, a Dutch island in the Caribbean with a population of about 159,000, became the smallest country to ever qualify for the World Cup tournament by using its own data and technology to “diaspora tracking”: mapping origins, identifying eligible players and using geospatial data to plan scouting trips and organize tryouts.
“Only one Curaçao 26 player was born on the island of Curaçao,” says Alex Stewart, chief executive of sports data consultancy Analytics FC. “The rest were born in the Netherlands.”
Another growing apply of data and AI in national federations is manager selection. The tools can analyze a pool of realistic squad options and identify managers whose tactical strengths best suit them. Teams can additionally apply AI to aid shape their pre-tournament lineup based on their group stage opponents.
England apply artificial intelligence to analyze penalties, knowing that penalties could knock them out. What used to take five days – analyzing every penalty taker against the opposition – can now probably be done in five hours, says the Football Association’s director of performance insights and analysis he told the BBC.
Marcelo Bielsa, the Uruguay manager, once said that when he was in charge of Premier League side Leeds United, his staff spent around 300 hours analyzing the upcoming team. “We can do it automatically,” says Lucey. It shows a video of red and blue dots moving around the pitch chasing a yellow ball. Analysts can ask questions – how often a given move resulted in shots or goals, and other times it did – each revealing a recent layer of information.
“Today, you can compare this situation with Internet access,” says Jan Wendt, co-founder and CEO of PLAIER, an AI platform cooperating with clubs and national teams. Both British Airways and Amazon created websites in the early days of the Internet. One became an information and airline ticketing platform, the other changed commerce around the world, Wendt says. Artificial intelligence has a similar reach, transforming both routine tasks and entire industries. Or, in the case of soccer, sports franchises.
However, AI tools and the staff required to build and operate them are exorbitant. Not all countries have the necessary resources. Wendt believes that working with already existing external companies, such as his own, should be seen as a more effective option for smaller countries.
Another challenge is that more data can make the analyst’s job more tough. Their role is to transform huge amounts of information into a handful of useful insights for the coach or player.
“You don’t want to say, ‘OK, now we can use all this cool stuff, here’s a 47-page document on your opposing full-back,’” Stewart says. “The analyst’s job is easier in some ways because there is more information. But it is more difficult because there is more information, so the ability to boil it down is required.”
The technology can improve match analysis and preparation for teams that previously couldn’t compete with countries with enormous scouting and analytics departments. But does this mean that the problem now will be that they won’t be able to compete with enormous teams of IT and analysts?
Will the data gap between richer and poorer countries widen so significantly as to disrupt competition in a tournament where the odds against smaller nations are already equal?
FIFA is concerned enough to have created a customized artificial intelligence agent, Football AI Pro, and is making it available to every country participating in the World Cup for the first time during this tournament.
The agent resembles a ChatGPT-style interface where coaches can enter questions and unlock information about their next opponents. Matches are played in 3D, which allows for analysis from previously impossible angles. All is measurable, from where players pass and run, through the way they attack and defend, to shots taken and goals scored.
“We consider it our goal, or even task, to provide technology to all teams, so that everyone has access to it and can use it easily, without the need to have additional experts in the team, because not everyone can afford it,” says Johannes Holzmüller, FIFA’s director of innovation.
It is debatable whether this will close the gap between a nation with a core data department and, say, an England national team that employs in-house programmers, data scientists and analysts using external AI tools. “This is the least we can do,” adds Holzmüller. “We see this gap when some teams use technology and data more than others.”
The future of data, artificial intelligence and football is, quite literally, about predicting the future. “The next step is long-term forecasting,” Lucey says, believing they will reach a point where counterfactual analysis will enable them to recommend players rest to maximize the likelihood of success.
Will FIFA have to step in and restrict nations to only using FIFA-approved AI tools?
“This is an important question,” says Holzmüller. “If it is regulated in some way then, there is no answer for today, but [AI] will play a big role in the future.”